Prosecution Insights
Last updated: November 04, 2026
Application No. 18/901,588

ANOMALY DETECTION FOR IDENTIFYING EXPOSURE EVENTS FROM BASELINE MOLECULAR MEASUREMENTS IN HUMAN HEALTH

Non-Final OA §103
Filed
Sep 30, 2024
Priority
Dec 08, 2023 — provisional 63/607,623
Examiner
SZUMNY, JONATHON A
Art Unit
3686
Tech Center
3600 — Transportation & Electronic Commerce
Assignee
Central Intelligence Agency
OA Round
3 (Non-Final)
57%
Grant Probability
Moderate
3-4
OA Rounds
9m
Est. Remaining
99%
With Interview

Examiner Intelligence

Grants 57% of resolved cases
57%
Career Allowance Rate
159 granted / 279 resolved
+5.0% vs TC avg
Strong +55% interview lift
Without
With
+55.1%
Interview Lift
resolved cases with interview
Typical timeline
2y 11m
Avg Prosecution
42 currently pending
Career history
336
Total Applications
across all art units

Statute-Specific Performance

§101
32.2%
-7.8% vs TC avg
§103
31.7%
-8.3% vs TC avg
§102
10.0%
-30.0% vs TC avg
§112
21.9%
-18.1% vs TC avg
Black line = Tech Center average estimate • Based on career data from 279 resolved cases

Office Action

§103
Notice of Pre-AIA or AIA Status The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA . Continued Examination Under 37 CFR 1.114 A request for continued examination under 37 CFR 1.114 (“RCE”), including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on July 13, 2026, has been entered. Status of Claims Claims 1-20 were previously pending and subject to a Final Office Action having a notification date of April 23, 2026 (“Final Office Action”). Following the Final Office Action, Applicant filed the RCE and an amendment on July 13, 2026 (“Amendment”), amending claims 1, 9, and 17. The present non-final Office Action addresses pending claims 1-20 in the Amendment. Response to Arguments Response to Applicant’s Arguments Regarding Claim Rejections Under 35 USC §102/103 Applicant’s arguments are moot in view of the new grounds of rejection as necessitated by the Amendment. Claim Rejections - 35 USC § 103 In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status. The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action: A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made. The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows: 1. Determining the scope and contents of the prior art. 2. Ascertaining the differences between the prior art and the claims at issue. 3. Resolving the level of ordinary skill in the pertinent art. 4. Considering objective evidence present in the application indicating obviousness or nonobviousness. This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention. Claims 1, 2, 4, 6-10, 12, 14-18, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Patent App. Pub. No. 2021/0358626 to Nicula et al. ("Nicula") in view of U.S. Patent App. Pub. No. 2024/0241094 to Remes et al. (“Remes”): Regarding claim 1, Nicula discloses a computer implemented method for processing an unlabeled omics feature vector ([0345]-[0346] discloses using a trained autoencoder to determine whether methylation patterns/stages in CpG sites (omics features which are vectors per [0094]-[0095], [0380], [0384]; also, [0088]-[0098] discloses various "omics" data) of a test subject correspond to a cancer state; furthermore, the omics features of the test subject are "unlabeled" as [0074], [0196] disclose the test samples are "unknown," hence the reason for using the autoencoder to determine the cancer state of the test subject in the first place), comprising: receiving, by one or more processors, a plurality of datasets containing omics feature vectors ([0339] discloses obtaining a training dataset including methylation patterns/fragments and corresponding nucleic acid sequences including methylation states of CpG sites (omics feature vectors) per [0094]-[0095], [0380], [0384]; for instance, Figure 1 and [0114]-[0115] illustrate/disclose various training data 124-130 for each of a plurality of training subjects 122; therefore, the various training data for each subject is a "dataset containing omics feature vectors" such that there are "a plurality of datasets containing omics feature vectors"; furthermore, processing core 102 in Figure 1 and [0053] disclose processor(s)), …; preprocessing, by the one or more processors, the plurality of datasets to produce a preprocessed dataset ([0158], [0194], [0198], [0201] discuss performing preprocessing/filtering the training datasets of each of the subjects (which would result in a "preprocessed dataset") before training of the autoencoder), wherein the preprocessing comprises: removing … technical noise across the plurality of datasets using one or more normalization techniques to produce a corresponding plurality of denoised datasets ([0158], [0167]-[0179] discloses how the filtering (preprocessing as noted above) can include analyzing and removing certain one of the methylation/nucleic acid fragments in the training datasets using various techniques ("normalization techniques") such as when they have less than a threshold sequencing depth, number of CpG sites, number of residues; are duplicates; also, [0194]-[0195] discloses how the preprocessing can include filtering by size, splitting, truncating, padding, etc. (all of which is removing "technical noise"); also, [0181] discloses removing noisy fragments); removal of the "technical noise" would result in "denoised datasets"); and aggregating the plurality of denoised datasets to generate the preprocessed dataset ([0198] discloses using the preprocessed/filtered fragments 124 ("denoised datasets") to train the autoencoder, where the "denoised datasets" of the subjects collectively represent an aggregated "preprocessed dataset"); training, by the one or more processors, a machine learning model using a subset of preprocessed omics feature vectors labeled as non-anomalous to perform anomaly detection ([0338]-[0340] discloses training an autoencoder using the training dataset (which can previously be preprocessed/filtered as noted above per [0167]-[0181], [0194]-[0195] resulting in the preprocessed dataset as noted above) to detect a cancer state (perform anomaly detection), where the preprocessed training dataset includes methylation patterns/fragments and corresponding nucleic acid sequences including methylation states of CpG sites (omics features) which are vectors per [0094]-[0095], [0380], [0384]; furthermore, as the end of [0339] notes how the training subjects have the cancer state which can be the absence of cancer per [0350], then some subset of the preprocessed omics feature vectors are labeled not having cancer (non-anomalous); providing, by the one or more processors, an unlabeled omics feature vector to the trained machine learning model ([0346] discloses applying the trained autoencoder to a test dataset including methylation/nucleic acid fragments including methylation states of CpG sites (omics features) which are vectors per [0094]-[0095], [0380], [0384]; furthermore, the omics feature vectors of the test subject are "unlabeled" as [0074], [0196] disclose how the test samples are "unknown," hence the reason for using the autoencoder to determine the cancer state of the test subject in the first place); generating, using the trained machine learning model, a low-dimensional latent space representation of the unlabeled omics feature vector ([0206] discloses how the autoencoder encodes a compressed (low-dimensional) latent representation of the test input (the unlabeled omics feature vector)); reconstructing, using the trained machine learning model, the unlabeled omics feature vector from the low-dimensional latent space representation to produce a reconstructed feature vector ([0206], [0346] disclose using the compressed latent representation to reconstruct the unknown test input (reconstructing the unlabeled omics feature vector to produce a "reconstructed feature vector")); evaluating a reconstruction error between the unlabeled omics feature vector and the reconstructed feature vector against an anomaly threshold to obtain a reconstruction error evaluation ([0346] discloses computing a score based on the reconstruction while [0050] discloses how the score is a loss/error between the reconstructed and inputted fragments (feature vectors) and [0281]-[0284] discloses how the score/loss/error is evaluated against a threshold); generating a feature label indicating whether the unlabeled omics feature vector is anomalous or non-anomalous based on the reconstruction error evaluation ([0281]-[0283], [0291], [0346] disclose how the fragment is indicative of the cancer state (which can be presence (anomalous) or absence (non-anomalous) of cancer per [0080], [0085], [0342], [0349]-[0350]) based on the evaluation of the score in view of the threshold; furthermore, the output determination of the cancer state by the trained autoencoder (e.g., see end of claim 135) is a generated feature label); and assigning the feature label to the unlabeled omics feature vector as metadata ([0326]-[0327] discloses how the cancer state determination can be repeated over time for continual cancer monitoring/assessments; accordingly, the feature label would be assigned to the omics feature vector of the test subject at each respective time to allow for such monitoring/assessments (e.g., to allow medical professionals to assess any similarities/differences in the omics feature vectors leading to the cancer state determinations over time); still further, the feature label would be assigned as "metadata" because the feature label is data about data (the later data being the omics feature vectors)). However, Nicula appears to be silent regarding each dataset of the plurality of datasets being from a different study, each study being associated with a different experimental design or execution. Nevertheless, Remes teaches ([0078]-[0079], [0117]-[0119]) that it was known in the machine learning and healthcare informatics art to train an ML model (e.g., autoencoder) based on a plurality of datasets of training data (e.g., biological components per [0031]) acquired during multiple different experiments/studies performed under different sets of experiment conditions that are performed under different sets of experiment conditions (different experiment design/executions) to advantageously enable use of the ML model across a wide range of experiment conditions. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention for each dataset of the plurality of datasets of Nicula to be from a different study, each study being associated with a different experimental design or execution, similar to as taught by Remes to advantageously enable use of the ML model across a wide range of experiment conditions. A person of ordinary skill in the art would have been motivated to combine the prior art to achieve the claimed invention and there would have been a reasonable expectation of success in doing so. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007). Furthermore, all the claimed elements were known in the prior art and one skilled in the art could have combined the elements as claimed by known methods with no change in their respective functions, and the combination yielded nothing more than predictable results to one of ordinary skill in the art. Id. As Nicula already discloses ([0158], [0167]-[0179], [0194], [0198], [0201]) preprocessing/filtering the datasets (which are from different studies associated with different experimental designs/executions per the above combination with Remes) by removing technical noise across the datasets using normalization techniques to produce denoised datasets as discussed above, then the removed technical noise is “inter-study” technical noise. Regarding claim 2, the Nicula/Remes combination discloses the computer implemented method of claim 1, further including wherein the preprocessing further comprises: labeling the omics feature vectors in each of the plurality of datasets as anomalous or non-anomalous ([0114] of Nicula discloses how the first training subjects (represented by omics feature vectors per [0339], [0094]-[0095], [0380], [0384] as noted previously) have the first cancer state (e.g., healthy/non-cancer/non-anomalous per [0128]) and second training subjects have the second cancer state (e.g., cancer/anomalous per [0128]); in order to distinguish the first and second training subjects, they would necessarily be labeled to indicate the respective first/second cancer state); and separating the preprocessed dataset into the subset of preprocessed omics feature vectors labeled as non-anomalous ([0158], [0194], [0198], [0201] of Nicula discloses how the first training dataset (labeled as healthy/non-cancer/non-anomalous as noted above) can be filtered/preprocessed which results in a subset of preprocessed omics feature vectors labeled as non-anomalous) and a subset of preprocessed omics feature vectors labeled as anomalous ([0238]-[0239] of Nicula discloses how the second training dataset (labeled as cancer/anomalous as noted above) can be filtered/preprocessed which results in a separate subset of preprocessed omics feature vectors labeled as non-anomalous). Regarding claim 4, the Nicula/Remes combination discloses the computer implemented method of claim 1, further including wherein the subset of preprocessed omics feature vectors labeled as non-anomalous comprises the set of omics feature vectors across each of the plurality of denoised datasets that are identified to not be associated with a disease, an illness, or adverse health symptoms ([0114] of Nicula discloses how the first training subjects (represented by omics feature vectors per [0339], [0094]-[0095], [0380], [0384] as noted previously) and which are denoised per [0158], [0167]-[0179] as discussed previously have the first cancer state (e.g., healthy/non-cancer per [0128]) (not associated with disease/illness/adverse health symptoms). Regarding claim 6, the Nicula/Remes combination discloses the computer implemented method of claim 1, further including wherein the omics feature vectors comprise gene expression levels or methylation statuses of mRNA, miRNA, methylated DNA, or microbiomes ([0339] of Nicula discloses obtaining a training dataset including methylation fragments/patterns and corresponding nucleic acid sequences including methylation states/statuses of CpG sites (omics features) which are vectors per [0094]-[0095], [0380], [0384]; also, [0045] discloses targeted DNA methylation sequencing, [0093] discloses methylated DNA fragments, and [0096] discloses DNA methylation). Regarding claim 7, the Nicula/Remes combination discloses the computer implemented method of claim 1, further including wherein the machine learning model is an autoencoder or a convolutional neural network ([0338]-[0340] of Nicula discloses training an autoencoder). Regarding claim 8, the Nicula/Remes combination discloses the computer implemented method of claim 1, further including wherein the plurality of datasets are public repository datasets or generated datasets ([0137]-[0154] of Nicula discloses performing one or more types of sequencing of biological samples to obtain the methylation/nucleic acid patterns/fragments (the datasets)). Regarding claim 9, Nicula discloses a system (system 100 in Figure 1), comprising: one or more memories (memory 111 in Figure 1); at least one processor (processing core 102 in Figure 1) each coupled to at least one of the memories and configured to perform operations. The remaining limitations of claim 9 are disclosed by the Nicula/Remes combination as discussed above in relation to claim 1. Claims 10, 12, and 14-16 are rejected in view of the Nicula/Remes combination as respectively discussed above in relation to claims 2, 4, and 6-8. Claims 17, 18, and 20 are rejected in view of the Nicula/Remes combination as respectively discussed above in relation to claims 9, 10, and 12. Claims 3, 11, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Patent App. Pub. No. 2021/0358626 to Nicula et al. ("Nicula") in view of U.S. Patent App. Pub. No. 2021/0257047 to Zhu et al. ("Zhu"): Regarding claim 3, the Nicula/Remes combination discloses the computer implemented method of claim 1, further including wherein the preprocessing further comprises: …using principal component analysis (the end of [0105] of Nicula discloses obtaining principal components of the training dataset which would necessarily involve use of PCA). However, the Nicula/Remes combination appears to be silent regarding use of such PCA confirms that the technical noise across the plurality of datasets has been removed. Nevertheless, Zhu teaches ([0421]-[0423]) that it was known in the healthcare informatics art to utilize PCA to remove noise from patient samples (including omics data per [0024]) to effectively improve normalization thereby increasing the accuracy of downstream analyses. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention for the PCA of the Nicula/Remes combination to confirm that the technical noise across the datasets has been removed similar to as taught by Zhu to effectively improve normalization thereby increasing the accuracy of downstream analyses. A person of ordinary skill in the art would have been motivated to combine the prior art to achieve the claimed invention and there would have been a reasonable expectation of success in doing so. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007). Furthermore, all the claimed elements were known in the prior art and one skilled in the art could have combined the elements as claimed by known methods with no change in their respective functions, and the combination yielded nothing more than predictable results to one of ordinary skill in the art. Id. Claims 11 and 19 are rejected in view of the Nicula/Remes/Zhu combination as discussed above in relation to claim 3. Claims 5 and 13 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Patent App. Pub. No. 2021/0358626 to Nicula et al. ("Nicula") in view of U.S. Patent App. Pub. No. 2025/0022542 to Poole et al. ("Poole"): Regarding claim 5, the Nicula/Remes combination discloses the computer implemented method of claim 1 but appears to be silent regarding wherein the one or more normalization techniques comprise quantile normalization, surrogate variable estimation, or z-score normalization, or combinations thereof. Nevertheless, Poole teaches ([0074]) that it was known in the healthcare informatics art to Z-score normalize omics data to facilitate determination of overexpression and underexpression of biomolecules thereby improving downstream analyses such as colon cancer screening and the like ([0007]). Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention for the one or more normalization techniques of the Nicula/Remes combination to include z-score normalization similar to as taught by Poole to advantageously facilitate determination of overexpression and underexpression of biomolecules thereby improving downstream analyses such as colon cancer screening and the like. person of ordinary skill in the art would have been motivated to combine the prior art to achieve the claimed invention and there would have been a reasonable expectation of success in doing so. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007). Furthermore, all the claimed elements were known in the prior art and one skilled in the art could have combined the elements as claimed by known methods with no change in their respective functions, and the combination yielded nothing more than predictable results to one of ordinary skill in the art. Id. Claim 13 is rejected in view of the Nicula/Remes/Poole combination as discussed above in relation to claim 5. Conclusion Any inquiry concerning this communication or earlier communications from the examiner should be directed to JONATHON A. SZUMNY whose telephone number is (303) 297-4376. The examiner can normally be reached Monday-Friday 7-5. Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice. If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Jason Dunham, can be reached at 571-272-8109. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300. Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000. /JONATHON A. SZUMNY/Primary Examiner, Art Unit 3686
Read full office action

Prosecution Timeline

Show 2 earlier events
Apr 02, 2026
Response Filed
Apr 23, 2026
Final Rejection mailed — §103
May 26, 2026
Interview Requested
Jun 04, 2026
Applicant Interview (Telephonic)
Jun 04, 2026
Examiner Interview Summary
Jul 13, 2026
Request for Continued Examination
Jul 20, 2026
Response after Non-Final Action
Aug 27, 2026
Non-Final Rejection mailed — §103 (current)

Precedent Cases

Applications granted by this same examiner with similar technology

Patent 12776409
EAR-WEARABLE DEVICE SYSTEMS AND METHODS FOR MONITORING EMOTIONAL STATE
4y 11m to grant Granted Oct 27, 2026
Patent 12755850
Systems and Methods to Generate a Surgical Risk Score and Uses Thereof
3y 0m to grant Granted Oct 06, 2026
Patent 12744125
Personal Profile Generator and Recommendation Engine
2y 0m to grant Granted Sep 22, 2026
Patent 12737618
SUPERCLASS-CONDITIONAL GAUSSIAN MIXTURE MODEL FOR PERSONALIZED PREDICTION ON DIALYSIS EVENTS
3y 11m to grant Granted Sep 15, 2026
Patent 12697070
WEARABLE HEALTH MONITORS AND METHODS OF MONITORING HEALTH
2y 9m to grant Granted Aug 04, 2026
Study what changed to get past this examiner. Based on 5 most recent grants.

Strategy Recommendation AI-generated — please review before filing

Get a prosecution strategy drawn from examiner precedents, rejection analysis, and claim mapping.
Typically takes 5-10 seconds — AI-generated, attorney review required before filing

Prosecution Projections

3-4
Expected OA Rounds
57%
Grant Probability
99%
With Interview (+55.1%)
2y 11m (~9m remaining)
Median Time to Grant
High
PTA Risk
Based on 279 resolved cases by this examiner. Grant probability derived from career allowance rate.

Sign in with your work email

Enter your email to receive a magic link. No password needed.

Personal email addresses (Gmail, Yahoo, etc.) are not accepted.

Free tier: 3 strategy analyses per month